INDEX

Explanations

words related to or associated with names

oai_token-act-pair · gpt-3.5-turbo

references to individuals named John

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤ©

-0.66

ï¸ı

-0.64

PF

-0.62

 course

-0.60

ãĤ°

-0.59

Statement

-0.59

 merit

-0.58

âĢ¢âĢ¢

-0.58

MET

-0.57

Limit

-0.55

POSITIVE LOGITS

anchester

0.79

ighty

0.68

 Kennedy

0.67

 Reilly

0.67

 Sabha

0.65

uez

0.64

ullah

0.64

aldo

0.64

 wiser

0.64

anton

0.63

Activations Density 0.096%

words related to or associated with names

references to individuals named John

No Comments

No Known Activations